INDEX

Explanations

phrases indicating inclusion or association with a group or concept

oai_token-act-pair · gpt-3.5-turbo

phrases emphasizing participation and belonging within communities or groups

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 assumes

-0.72

ptive

-0.64

oa

-0.63

ants

-0.61

 directs

-0.60

 withd

-0.60

 opposes

-0.59

 relies

-0.58

eg

-0.57

culosis

-0.55

POSITIVE LOGITS

ĺħ

0.79

 something

0.77

 history

0.72

İĭ

0.72

 society

0.71

 folklore

0.71

Ļ½

0.69

¬¼

0.68

 Kinnikuman

0.63

history

0.63

Activations Density 0.088%

phrases indicating inclusion or association with a group or concept

phrases emphasizing participation and belonging within communities or groups

No Comments

No Known Activations