INDEX

Explanations

words related to causation or consequences

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs49152-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_49152

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

torch.float32

Hook Point

blocks.8.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

 rehabilit

-0.69

 hitter

-0.68

abies

-0.60

 veterinarian

-0.58

 nurs

-0.57

 batter

-0.57

 neighb

-0.56

 territ

-0.55

 battered

-0.53

igger

-0.53

POSITIVE LOGITS

forth

2.24

forward

1.44

why

1.01

why

0.89

noon

0.80

entimes

0.80

ably

0.80

videos

0.77

fter

0.76

xual

0.73

Activations Density 0.009%

words related to causation or consequences

No Comments

No Known Activations