INDEX

Explanations

phrases related to political actions and organizations

oai_token-act-pair · gpt-3.5-turbo

suffixes and word endings that suggest traits or characteristics

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 endowed

-0.64

enhagen

-0.63

 sterling

-0.62

 Independence

-0.62

 Pavilion

-0.61

 Calder

-0.60

 Kessler

-0.60

ACTED

-0.58

 Shotgun

-0.58

 Carnage

-0.57

POSITIVE LOGITS

adic

0.80

oslav

0.77

pta

0.77

egu

0.76

oby

0.76

omon

0.75

nit

0.74

omi

0.74

ivas

0.74

omore

0.73

Activations Density 0.085%

phrases related to political actions and organizations

suffixes and word endings that suggest traits or characteristics

No Known Activations