INDEX

Explanations

text patterns that present two alternative possibilities or contrasting statements, typically starting with "Or"

oai_token-act-pair · gpt-3.5-turbo

the phrase "Or" frequently, indicating a focus on presenting alternatives or different perspectives

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

owner

-0.77

IDENT

-0.71

Material

-0.66

makers

-0.64

encers

-0.63

ï¿½ï¿½ï¿½

-0.63

ailable

-0.63

KER

-0.61

UD

-0.61

NOW

-0.61

POSITIVE LOGITS

thodox

1.30

lando

1.20

phans

1.20

chard

1.17

Else

1.15

phan

1.10

 alternatively

1.05

acles

0.98

anges

0.98

leans

0.97

Activations Density 0.044%

text patterns that present two alternative possibilities or contrasting statements, typically starting with "Or"

the phrase "Or" frequently, indicating a focus on presenting alternatives or different perspectives

No Comments

No Known Activations