INDEX

Explanations

the word "reason" and phrases related to justifying or explaining actions or events

oai_token-act-pair · gpt-3.5-turbo

phrases that indicate the causes or reasons behind events or situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.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

ãĤ´ãĥ³

-0.70

KY

-0.69

chron

-0.66

 Carbuncle

-0.64

boro

-0.62

 redes

-0.62

 livest

-0.62

borg

-0.62

aith

-0.61

agog

-0.61

POSITIVE LOGITS

why

1.60

abl

1.27

why

1.26

WHY

1.14

 behind

1.09

Why

0.92

Why

0.90

 cited

0.79

 cele

0.74

 rationale

0.72

Activations Density 0.042%

the word "reason" and phrases related to justifying or explaining actions or events

phrases that indicate the causes or reasons behind events or situations

No Comments

No Known Activations