INDEX

Explanations

the word 'Iron' followed by another word

oai_token-act-pair · gpt-3.5-turbo

mentions of "Iron" in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

uated

-0.81

arre

-0.75

enance

-0.74

soType

-0.72

CLS

-0.66

ortion

-0.66

itia

-0.66

ired

-0.65

igate

-0.64

uates

-0.64

POSITIVE LOGITS

clad

1.14

ore

0.91

axe

0.86

 marrow

0.84

 Iron

0.83

Age

0.81

works

0.80

forge

0.79

mong

0.78

claw

0.76

Activations Density 0.007%

the word 'Iron' followed by another word

mentions of "Iron" in various contexts

No Comments

No Known Activations