INDEX

Explanations

tokens related to personal or proper names

oai_token-act-pair · gpt-3.5-turbo

proper nouns or names of people

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

nered

-0.95

lain

-0.90

rug

-0.89

hof

-0.84

agher

-0.83

lich

-0.80

rums

-0.78

role

-0.78

rarily

-0.78

furt

-0.78

POSITIVE LOGITS

sie

0.77

lectic

0.66

IVER

0.63

 payday

0.61

ESSION

0.61

ox

0.59

LECT

0.59

FINEST

0.58

 opioids

0.58

 Prescott

0.56

Activations Density 0.161%

tokens related to personal or proper names

proper nouns or names of people

No Comments

No Known Activations