INDEX

Explanations

phrases related to paraphrasing

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "paragraph" and related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

DERR

-0.77

ships

-0.77

Kit

-0.77

eer

-0.76

 Dragonbound

-0.76

 Lauder

-0.75

Dash

-0.72

Beh

-0.72

 Lilly

-0.71

 Shine

-0.71

POSITIVE LOGITS

 paraph

1.08

igm

0.99

asms

0.90

antom

0.89

":["

0.85

inct

0.85

iscal

0.85

orne

0.84

odied

0.82

ropolitan

0.81

Activations Density 0.018%

phrases related to paraphrasing

occurrences of the word "paragraph" and related terms

No Comments

No Known Activations