INDEX

Explanations

online content-related items or actions

oai_token-act-pair · gpt-3.5-turbo

proper formatting and structure in a document, especially related to categorizations and classifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 citiz

-0.73

 tremend

-0.72

quartered

-0.62

 Skydragon

-0.61

 footing

-0.61

 aging

-0.58

 Younger

-0.57

cius

-0.56

omething

-0.56

 spokes

-0.55

POSITIVE LOGITS

↵

0.77

Ì¶

0.68

ÂłÂłÂł

0.63

iona

0.62

Í

0.62

à¦

0.62

dding

0.62

âĵĺ

0.61

Merit

0.61

gex

0.60

Activations Density 0.174%

online content-related items or actions

proper formatting and structure in a document, especially related to categorizations and classifications

No Comments

No Known Activations