INDEX

Explanations

words related to discussions, opinions, and activities within specific communities

oai_token-act-pair · gpt-3.5-turbo

references to communities or specific groups

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

blem

-0.77

Weapons

-0.74

Lua

-0.67

falls

-0.64

anas

-0.64

errors

-0.64

Christmas

-0.64

 whilst

-0.62

 valves

-0.61

Shares

-0.61

POSITIVE LOGITS

 vicinity

1.24

 audience

1.22

 community

0.95

 region

0.90

 neighborhood

0.87

 area

0.86

 profession

0.85

 workforce

0.85

 privileged

0.85

 field

0.84

Activations Density 0.141%

words related to discussions, opinions, and activities within specific communities

references to communities or specific groups

No Comments

No Known Activations