INDEX

Explanations

search terms related to law enforcement techniques and statistics

oai_token-act-pair · gpt-3.5-turbo

contexts related to quantitative comparisons and statistics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

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

urai

-0.60

 groove

-0.56

 reluctantly

-0.56

asma

-0.55

 alias

-0.54

 burner

-0.52

 ASAP

-0.52

 somew

-0.52

 thanking

-0.52

anta

-0.52

POSITIVE LOGITS

 predecessors

0.99

 elsewhere

0.84

 counterparts

0.83

 previous

0.82

 predec

0.80

 preceding

0.78

 equivalents

0.77

 predecessor

0.76

 conventional

0.74

 comparable

0.73

Activations Density 0.784%

search terms related to law enforcement techniques and statistics

contexts related to quantitative comparisons and statistics

No Comments

No Known Activations