INDEX

Explanations

phrases related to methods or processes

oai_token-act-pair · gpt-3.5-turbo

mentions of various methods

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

 impression

-0.64

 Aliens

-0.61

minster

-0.60

ments

-0.60

 gloom

-0.60

 Falcons

-0.60

 Sons

-0.60

 Barkley

-0.59

igi

-0.59

 Germ

-0.57

POSITIVE LOGITS

ologies

1.44

ologic

1.09

ologically

1.08

OLOGY

0.95

ical

0.95

ology

0.95

sonian

0.89

ically

0.86

fare

0.85

Override

0.84

Activations Density 0.020%

phrases related to methods or processes

mentions of various methods

No Comments

No Known Activations