INDEX

Explanations

names of certain individuals or companies

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and entities in a public or political context

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

hed

-0.79

bledon

-0.79

ftime

-0.78

ufact

-0.77

idated

-0.75

unts

-0.72

tesy

-0.70

oppable

-0.70

ascus

-0.70

pering

-0.70

POSITIVE LOGITS

lev

0.72

lect

0.68

 capsules

0.65

 Sector

0.65

 Pred

0.60

NES

0.60

 Committees

0.60

LM

0.59

emouth

0.59

 Liberty

0.59

Activations Density 0.036%

names of certain individuals or companies

references to specific individuals and entities in a public or political context

No Comments

No Known Activations