INDEX

Explanations

the interjection "um," which indicates hesitation or uncertainty in speech

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

 Dayton

-0.69

 Reborn

-0.65

ttes

-0.64

 Stevenson

-0.63

Kings

-0.62

hips

-0.62

%%

-0.61

 undergone

-0.61

çĶŁ

-0.59

ingham

-0.58

POSITIVE LOGITS

pty

1.15

ming

1.03

br

0.98

ihara

0.96

cond

0.93

mers

0.91

ami

0.89

pling

0.88

pt

0.88

amoto

0.88

Activations Density 0.003%

the interjection "um," which indicates hesitation or uncertainty in speech

No Comments

No Known Activations