INDEX

Explanations

the repeated use of the word "and."

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

ufficient

-0.68

rice

-0.66

catentry

-0.63

iment

-0.61

iction

-0.60

NULL

-0.58

ongo

-0.58

panel

-0.58

 Older

-0.56

jer

-0.56

POSITIVE LOGITS

rogen

0.83

romeda

0.78

 explore

0.76

uin

0.74

rew

0.73

seek

0.72

 forth

0.71

 signings

0.68

roph

0.68

buy

0.68

Activations Density 0.042%

the repeated use of the word "and."

No Comments

No Known Activations