INDEX

Explanations

proper nouns or names, specifically the name "Malcolm"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Malcolm."

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

spring

-0.78

ptroller

-0.74

rums

-0.71

 payable

-0.70

sburgh

-0.65

 habitable

-0.65

mble

-0.64

arte

-0.63

ricular

-0.63

repre

-0.63

POSITIVE LOGITS

colm

1.16

 Turnbull

1.14

sey

0.98

 Malcolm

0.95

 Tucker

0.87

ces

0.77

 Curtis

0.77

 Glad

0.75

 Fraser

0.73

 Goodwin

0.72

Activations Density 0.005%

proper nouns or names, specifically the name "Malcolm"

mentions of the name "Malcolm."

No Comments

No Known Activations