INDEX

Explanations

opinions, beliefs, and claims expressed by individuals

oai_token-act-pair · gpt-3.5-turbo

assertions and beliefs that may be misleading or false

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rontal

-0.92

ktop

-0.91

ograp

-0.77

 teasp

-0.74

yna

-0.73

arthed

-0.73

 exting

-0.72

Lex

-0.72

srf

-0.70

Pg

-0.70

POSITIVE LOGITS

 otherwise

1.12

 causation

0.86

 somehow

0.81

 conservatives

0.78

 liberals

0.78

 ignorance

0.78

 Muslims

0.78

 Asians

0.77

 vaccines

0.77

 incompetence

0.77

Activations Density 0.327%

opinions, beliefs, and claims expressed by individuals

assertions and beliefs that may be misleading or false

No Comments

No Known Activations