INDEX

Explanations

mentions of specific groups of people or professionals (e.g., senators, miners) and some related actions

oai_token-act-pair · gpt-3.5-turbo

references to specific groups or professions related to politics and social issues

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

Minecraft

-0.70

fff

-0.67

orthy

-0.65

 Tropical

-0.63

ciation

-0.61

lite

-0.61

cens

-0.61

 Britann

-0.60

cgi

-0.59

Pak

-0.58

POSITIVE LOGITS

 reluct

0.68

 complain

0.68

 frustrations

0.63

 balk

0.63

 rallied

0.62

 levers

0.61

 penchant

0.61

 retali

0.61

 electing

0.61

 withdrawals

0.60

Activations Density 0.558%

mentions of specific groups of people or professionals (e.g., senators, miners) and some related actions

references to specific groups or professions related to politics and social issues

No Known Activations