INDEX

Explanations

the word "citizen"

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "citizen" and its variations

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

hiba

-1.11

mith

-0.93

heet

-0.85

cale

-0.85

ecause

-0.83

creen

-0.82

ermott

-0.80

ension

-0.79

orp

-0.76

ammers

-0.76

POSITIVE LOGITS

ry

1.01

 citizen

0.89

 Citizen

0.82

hood

0.79

 Kane

0.74

naire

0.73

ãĤ©

0.73

 uprising

0.68

fare

0.65

 Volks

0.63

Activations Density 0.015%

the word "citizen"

occurrences of the word "citizen" and its variations

No Comments

No Known Activations