INDEX

Explanations

names specifically "Simon"

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the name "Simon."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

ttes

-0.68

een

-0.67

ingo

-0.64

ITNESS

-0.61

ations

-0.59

laws

-0.57

reek

-0.57

tub

-0.56

tc

-0.56

utic

-0.55

POSITIVE LOGITS

 Says

0.92

etta

0.88

etti

0.85

 Gerr

0.84

ãĤ¨ãĥ«

0.76

ovic

0.73

displayText

0.73

itars

0.70

idis

0.70

 Fraser

0.70

Activations Density 0.037%

names specifically "Simon"

repeated mentions of the name "Simon."

No Comments

No Known Activations