INDEX

Explanations

dates and numbers

oai_token-act-pair · gpt-3.5-turbo

positive reviews or praise for books

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 eleph

-1.10

 tremend

-1.10

 occas

-0.92

aditional

-0.88

 metic

-0.88

 exha

-0.85

 citiz

-0.85

 undermin

-0.84

 hemor

-0.84

oun

-0.84

POSITIVE LOGITS

↵

0.98

ÂłÂłÂł

0.83

\":

0.82

Merit

0.70

Õ

0.68

ÂłÂłÂłÂłÂłÂłÂłÂł

0.68

ÂłÂłÂłÂł

0.67

tle

0.66

Ø§ÙĦ

0.66

ETF

0.65

Activations Density 0.404%

dates and numbers

positive reviews or praise for books

No Comments

No Known Activations