INDEX

Explanations

terms related to politics, specifically focusing on "IT" as the highest activation

oai_token-act-pair · gpt-3.5-turbo

topics related to political themes or issues

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

 retri

-0.60

 Marketable

-0.59

 worry

-0.59

 Thumbnails

-0.59

 marqu

-0.58

 subtitle

-0.58

 happ

-0.58

 brace

-0.57

ldom

-0.57

liest

-0.57

POSITIVE LOGITS

geist

0.91

opol

0.81

insk

0.78

é¾

0.77

okia

0.76

hens

0.75

ãĤ´ãĥ³

0.74

iah

0.74

nik

0.72

kov

0.71

Activations Density 0.112%

terms related to politics, specifically focusing on "IT" as the highest activation

topics related to political themes or issues

No Comments

No Known Activations