INDEX

Explanations

proper nouns that contain the word "Sag"

oai_token-act-pair · gpt-3.5-turbo

names of characters, especially prominently mentioned ones

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

xual

-0.83

theless

-0.77

ateral

-0.77

thren

-0.72

byter

-0.70

terday

-0.68

 doub

-0.67

IGHT

-0.67

ervatives

-0.64

ple

-0.64

POSITIVE LOGITS

raf

1.01

rams

0.99

tag

0.99

pak

0.97

tags

0.97

rod

0.95

gers

0.95

rils

0.93

ril

0.93

Sag

0.91

Activations Density 0.014%

proper nouns that contain the word "Sag"

names of characters, especially prominently mentioned ones

No Comments

No Known Activations