INDEX

Explanations

terms related to legal and official actions

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

cheon

-0.64

Gas

-0.62

 Editors

-0.59

 horm

-0.57

verages

-0.57

Ba

-0.57

Jet

-0.56

acists

-0.56

sei

-0.55

arine

-0.55

POSITIVE LOGITS

olate

0.75

 noises

0.69

 mistake

0.69

RANT

0.66

sense

0.66

 unilaterally

0.66

.–

0.66

 elsewhere

0.65

ylum

0.64

 payable

0.62

Activations Density 0.077%

terms related to legal and official actions

No Comments

No Known Activations