INDEX

Explanations

words ending in 'ing' preceded by another word or character

oai_token-act-pair · gpt-3.5-turbo

words ending in 'ing' related to various actions or processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 unequ

-0.69

 unim

-0.68

 certific

-0.64

#$#$

-0.62

 isolate

-0.60

 check

-0.60

chwitz

-0.60

 undet

-0.60

 Skydragon

-0.60

 Register

-0.59

POSITIVE LOGITS

atively

0.90

 aloud

0.84

gow

0.76

ially

0.76

eth

0.76

terday

0.75

agher

0.75

edly

0.75

Sov

0.74

LY

0.72

Activations Density 0.093%

words ending in 'ing' preceded by another word or character

words ending in 'ing' related to various actions or processes

No Comments

No Known Activations