INDEX

Explanations

phrases related to giving instructions or directives

oai_token-act-pair · gpt-3.5-turbo

phrases that instruct or prompt actions

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

 defe

-0.61

ean

-0.61

lik

-0.58

 prophes

-0.58

NESS

-0.57

iege

-0.55

 Lich

-0.54

ikawa

-0.54

ridge

-0.53

ode

-0.52

POSITIVE LOGITS

rid

1.14

TING

1.11

away

0.96

aways

0.89

 acquainted

0.79

tin

0.78

cloneembedreportprint

0.78

ãĥ³ãĤ¸

0.78

 Started

0.74

 Away

0.72

Activations Density 0.081%

phrases related to giving instructions or directives

phrases that instruct or prompt actions

No Comments

No Known Activations