INDEX

Explanations

statements of facts or findings

oai_token-act-pair · gpt-3.5-turbo

discussions surrounding advantages, myths, and important facts

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

esi

-0.82

ebook

-0.80

ebin

-0.72

aline

-0.70

sonian

-0.68

azo

-0.68

arez

-0.67

 Fiction

-0.66

arus

-0.66

xon

-0.64

POSITIVE LOGITS

 pertaining

0.78

viz

0.77

:-

0.75

 relating

0.73

 imaginable

0.72

 onlook

0.68

 peculiar

0.68

 creep

0.67

 including

0.66

 propositions

0.65

Activations Density 0.332%

statements of facts or findings

discussions surrounding advantages, myths, and important facts

No Comments

No Known Activations