INDEX

Explanations

names or terms related to famous people or characters

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names

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

ikuman

-0.61

akeru

-0.58

etheus

-0.58

Pwr

-0.55

arbon

-0.55

ishi

-0.55

illions

-0.55

oola

-0.54

ovie

-0.53

rative

-0.53

POSITIVE LOGITS

aternity

0.61

heit

0.56

pert

0.56

hers

0.55

ngth

0.54

beard

0.54

clave

0.53

ascript

0.53

cko

0.53

pants

0.52

Activations Density 1.519%

names or terms related to famous people or characters

proper nouns, particularly names

No Comments

No Known Activations