INDEX

Explanations

names of people and possibly places

oai_token-act-pair · gpt-3.5-turbo

proper nouns, particularly names of people and characters

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Eliot

-0.62

âĹ¼

-0.60

KT

-0.58

 Kirk

-0.57

gling

-0.56

 anonymity

-0.56

 Irving

-0.56

 Shattered

-0.56

hed

-0.55

 hawk

-0.55

POSITIVE LOGITS

rosso

0.87

iola

0.85

berto

0.85

ondo

0.84

aldi

0.84

udo

0.84

 Aless

0.82

arez

0.80

orio

0.80

Ã¡

0.79

Activations Density 0.111%

names of people and possibly places

proper nouns, particularly names of people and characters

No Comments

No Known Activations