INDEX

Explanations

conjunctions or words indicating causality between events

oai_token-act-pair · gpt-3.5-turbo

the word "so" and its variants, indicating a focus on conversational transitions or emphasis

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

•Causality

so + thus,therefore

No Comments

Negative Logits

 Modes

-0.65

Newsletter

-0.60

 Passage

-0.57

ansk

-0.56

 Â®

-0.56

icipated

-0.55

degree

-0.55

weights

-0.54

sha

-0.53

Aust

-0.52

POSITIVE LOGITS

oooo

1.31

ooo

1.28

oner

1.17

 yeah

1.11

bered

1.09

oooooooo

1.04

 much

0.99

 glad

0.98

 damn

0.98

othe

0.97

Activations Density 0.075%

conjunctions or words indicating causality between events

the word "so" and its variants, indicating a focus on conversational transitions or emphasis

No Comments

No Known Activations