INDEX

Explanations

Latin and foreign language words or phrases

oai_token-act-pair · gpt-3.5-turbo

occurrences of a specific character or symbol

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 mathemat

-0.79

 disadvant

-0.75

awaru

-0.72

 carbohyd

-0.71

 soph

-0.71

merce

-0.70

 contrace

-0.66

efficients

-0.64

 Wichita

-0.64

 scholarly

-0.63

POSITIVE LOGITS

ï¸ı

1.15

å¥³

0.91

å£«

0.84

éĩ

0.79

ï¸

0.77

 istg

0.76

çľ

0.76

tu

0.76

æľ

0.75

äº

0.75

Activations Density 0.316%

Latin and foreign language words or phrases

occurrences of a specific character or symbol

No Comments

No Known Activations