INDEX

Explanations

references to written texts and literary works

oai_token-act-pair · gpt-3.5-turbo

references to literature and literary texts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

twitch

-1.05

 Twitch

-0.75

eto

-0.73

HUD

-0.71

Exit

-0.70

 twitch

-0.70

Join

-0.67

Connector

-0.67

NYSE

-0.67

Tonight

-0.66

POSITIVE LOGITS

 authored

1.46

 written

1.42

 penned

1.39

 published

1.38

 manuscript

1.22

 edited

1.22

 printed

1.20

 reprinted

1.19

 circulated

1.18

published

1.16

Activations Density 0.425%

references to written texts and literary works

references to literature and literary texts

No Comments

No Known Activations