INDEX

Explanations

the word "phrase" followed by an intense activation value

oai_token-act-pair · gpt-3.5-turbo

recurring phrases or sentence structures

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

DERR

-0.84

ÄŁ

-0.83

©¶æ¥µ

-0.77

 Thro

-0.73

 Skydragon

-0.70

Adds

-0.68

fman

-0.68

Ka

-0.66

hari

-0.65

ntil

-0.63

POSITIVE LOGITS

phrase

1.07

ology

1.06

 phrases

1.03

 phrase

1.02

 uttered

0.87

terday

0.86

witz

0.82

mith

0.77

atre

0.74

eting

0.74

Activations Density 0.016%

the word "phrase" followed by an intense activation value

recurring phrases or sentence structures

No Comments

No Known Activations