INDEX

Explanations

instances where something is almost happening or being done

oai_token-act-pair · gpt-3.5-turbo

phrases expressing a sense of near or approximate experiences

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

oran

-0.78

ãĤ¸

-0.76

 Dynamics

-0.73

 UNITED

-0.68

alf

-0.65

nation

-0.65

 è£ıè¦ļéĨĴ

-0.65

oÄŁ

-0.65

Republic

-0.64

 Belt

-0.63

POSITIVE LOGITS

PsyNetMessage

0.73

 certainly

0.71

 forgot

0.70

 spoil

0.68

lex

0.66

finished

0.66

 exha

0.65

anova

0.65

stress

0.65

Dial

0.64

Activations Density 0.039%

instances where something is almost happening or being done

phrases expressing a sense of near or approximate experiences

No Comments

No Known Activations