INDEX

Explanations

the word "nights" with strong activations

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "nights."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 bloc

-0.72

 Swiss

-0.66

lda

-0.66

ression

-0.65

ific

-0.65

offic

-0.63

 Canary

-0.63

BN

-0.61

eering

-0.61

 Democrat

-0.61

POSITIVE LOGITS

creen

1.35

mith

1.25

pring

1.12

hift

1.08

uits

1.06

cape

1.04

poons

1.03

hips

1.01

hops

1.01

cale

1.01

Activations Density 0.031%

the word "nights" with strong activations

occurrences of the word "nights."

No Comments

No Known Activations