INDEX

Explanations

proper nouns, specifically names such as "Howard"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Howard."

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

asus

-0.76

yrim

-0.76

terday

-0.69

ccording

-0.68

erala

-0.68

ktop

-0.66

BOOK

-0.66

CVE

-0.65

tera

-0.65

ebin

-0.64

POSITIVE LOGITS

 Stern

1.09

 Howard

1.04

 Dean

0.87

 Ellis

0.86

 Schultz

0.85

 Webb

0.81

Howard

0.80

 Neville

0.79

 Hughes

0.78

 Buffett

0.77

Activations Density 0.007%

proper nouns, specifically names such as "Howard"

mentions of the name "Howard."

No Comments

No Known Activations