INDEX

Explanations

mentions of food and dining-related terms

oai_token-act-pair · gpt-3.5-turbo

references to food or dining options

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ultan

-0.85

icus

-0.82

orial

-0.78

acca

-0.78

iversal

-0.70

izations

-0.70

ieth

-0.70

ulating

-0.69

 interf

-0.69

ulkan

-0.69

POSITIVE LOGITS

 fare

1.08

 fares

1.01

ttes

0.94

fare

0.81

well

0.78

ptin

0.76

bill

0.74

rer

0.72

jit

0.72

ways

0.72

Activations Density 0.007%

mentions of food and dining-related terms

references to food or dining options

No Comments

No Known Activations