INDEX

Explanations

names of people

oai_token-act-pair · gpt-3.5-turbo

proper nouns, specifically names of individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

robe

-0.93

shaw

-0.84

ebook

-0.83

arer

-0.80

away

-0.80

eds

-0.77

agra

-0.76

our

-0.76

ally

-0.75

lay

-0.75

POSITIVE LOGITS

ĨĴ

0.88

xtap

0.82

 Felix

0.82

Reloaded

0.81

entanyl

0.81

£ı

0.79

 confir

0.79

etheless

0.76

paio

0.76

theless

0.75

Activations Density 0.019%

names of people

proper nouns, specifically names of individuals

No Comments

No Known Activations