INDEX

Explanations

written text in sentences that act as evidence or formal communication

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "written."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

abe

-0.90

nel

-0.85

MAL

-0.77

alo

-0.76

elli

-0.76

Ĭ±

-0.73

abil

-0.73

nels

-0.71

allows

-0.71

Magn

-0.70

POSITIVE LOGITS

 written

0.95

 penned

0.81

 breath

0.80

written

0.76

 typed

0.75

 essays

0.74

writing

0.74

 writing

0.74

 essay

0.73

itatively

0.71

Activations Density 0.031%

written text in sentences that act as evidence or formal communication

occurrences of the word "written."

No Comments

No Known Activations