INDEX

Explanations

hyphenated words expressing negative connotations

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs3072-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_3072

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

3,072

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

 reservation

-0.61

 Chop

-0.59

 vigilance

-0.59

Rav

-0.59

 Responsibility

-0.58

 induction

-0.58

 enrichment

-0.57

 Rivals

-0.57

 Mick

-0.57

 elimination

-0.56

POSITIVE LOGITS

called

1.28

ndra

0.98

sembly

0.84

far

0.84

worldly

0.82

mentioned

0.82

ciation

0.77

much

0.74

soon

0.73

cial

0.72

Activations Density 0.892%

hyphenated words expressing negative connotations

No Comments

No Known Activations