INDEX

Explanations

links or connections between different concepts, typically highlighted by the word "between."

oai_token-act-pair · gpt-3.5-turbo

phrases indicating the existence of correlations or associations between different concepts

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

fl

-0.78

OGR

-0.77

quished

-0.77

nell

-0.76

entimes

-0.75

tsky

-0.75

kl

-0.75

di

-0.73

gn

-0.72

kt

-0.71

POSITIVE LOGITS

 sexes

0.69

 disparate

0.67

 genders

0.64

 scarce

0.62

 them

0.60

 observable

0.60

two

0.60

 bean

0.59

 criminality

0.59

 STATS

0.59

Activations Density 0.032%

links or connections between different concepts, typically highlighted by the word "between."

phrases indicating the existence of correlations or associations between different concepts

No Comments

No Known Activations