INDEX

Explanations

the top-ranking items or entities

oai_token-act-pair · gpt-3.5-turbo

the word "top" in various contexts

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

 Consent

-0.69

 Gaul

-0.68

 Kath

-0.62

otive

-0.61

 consent

-0.61

ufact

-0.60

asca

-0.59

cci

-0.59

Cla

-0.59

 vomit

-0.58

POSITIVE LOGITS

most

1.05

ographical

1.03

ography

0.93

ographically

0.93

deck

0.93

 notch

0.89

 tier

0.88

eka

0.85

unders

0.82

 billing

0.82

Activations Density 0.034%

the top-ranking items or entities

the word "top" in various contexts

No Comments

No Known Activations