INDEX

Explanations

verbs related to intervention or interference

oai_token-act-pair · gpt-3.5-turbo

instances of the word "sid" or related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

quality

-0.72

 Samoa

-0.64

URI

-0.63

ibaba

-0.62

xit

-0.61

NCT

-0.61

older

-0.59

FOX

-0.58

grade

-0.58

STD

-0.58

POSITIVE LOGITS

sid

1.47

etr

0.82

stead

0.71

 comprom

0.69

 Phar

0.68

roid

0.65

terson

0.65

 nods

0.65

 Tanz

0.65

stract

0.65

Activations Density 0.006%

verbs related to intervention or interference

instances of the word "sid" or related terms

No Comments

No Known Activations