INDEX

Explanations

terms related to public utilities

oai_token-act-pair · gpt-3.5-turbo

references to utility services

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

eworld

-0.91

orne

-0.85

uten

-0.84

bring

-0.79

oning

-0.79

artisan

-0.78

eworthy

-0.77

alien

-0.74

oria

-0.73

thus

-0.73

POSITIVE LOGITS

 utility

0.95

 utilities

0.90

 belt

0.85

 Utilities

0.82

gency

0.79

 bills

0.79

stration

0.76

 subst

0.73

Pwr

0.72

 meter

0.72

Activations Density 0.024%

terms related to public utilities

references to utility services

No Comments

No Known Activations