INDEX

Explanations

phrases related to authors and directors

oai_token-act-pair · gpt-3.5-turbo

words with specific suffixes, particularly "-ant," "-ent," and "-iant."

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

keyes

-0.76

ModLoader

-0.69

usterity

-0.66

âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦âĢ¦

-0.63

boats

-0.62

Ô

-0.61

Trade

-0.61

 manageable

-0.61

 mainline

-0.61

Â·Â·

-0.59

POSITIVE LOGITS

opol

0.97

0.95

ensen

0.91

otal

0.90

ais

0.88

opoulos

0.86

ea

0.83

nik

0.82

iy

0.82

inis

0.82

Activations Density 0.218%

phrases related to authors and directors

words with specific suffixes, particularly "-ant," "-ent," and "-iant."

No Comments

No Known Activations