INDEX

Explanations

references to sizes or measurements

oai_token-act-pair · gpt-3.5-turbo

phrases indicating strong emphasis or importance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

Political

-0.69

 Establishment

-0.66

 extremism

-0.65

 political

-0.64

 uncond

-0.64

political

-0.63

 ideological

-0.61

 DACA

-0.61

 substantive

-0.60

 membership

-0.59

POSITIVE LOGITS

 latex

0.65

 bladder

0.63

coni

0.63

 waterproof

0.62

 coated

0.61

 lubric

0.60

 smell

0.60

 wearer

0.60

chy

0.58

 pinch

0.58

Activations Density 2.042%

references to sizes or measurements

phrases indicating strong emphasis or importance

No Comments

No Known Activations