INDEX

Explanations

names of political figures and organizations

oai_token-act-pair · gpt-3.5-turbo

references to specific political entities and notable locations

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

minecraft

-0.57

lihood

-0.53

entary

-0.50

riv

-0.50

////

-0.49

obyl

-0.48

cellent

-0.47

bleacher

-0.47

abiding

-0.47

Marginal

-0.46

POSITIVE LOGITS

 treasury

0.63

 brass

0.55

 Ethics

0.53

 electorate

0.52

senal

0.52

Dialogue

0.52

 insiders

0.51

 Treasury

0.51

 payroll

0.51

 Citadel

0.50

Activations Density 0.952%

names of political figures and organizations

references to specific political entities and notable locations

No Comments

No Known Activations