INDEX

Explanations

words related to forms or formalities

oai_token-act-pair · gpt-3.5-turbo

references to different forms or structures

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

 Seas

-0.62

EStreamFrame

-0.61

 Regional

-0.61

 Dodgers

-0.60

 pronounced

-0.60

 paras

-0.60

po

-0.58

Sea

-0.58

 blazing

-0.58

 TOTAL

-0.58

POSITIVE LOGITS

aldehyde

1.28

idable

1.28

form

1.25

forms

1.06

ative

1.05

atted

1.03

ulas

0.98

ula

0.97

furt

0.96

ulations

0.93

Activations Density 0.006%

words related to forms or formalities

references to different forms or structures

No Comments

No Known Activations