INDEX

Explanations

phrases indicating a change or event occurring in a sequence

oai_token-act-pair · gpt-3.5-turbo

instances of the word "when" used to indicate time or conditional scenarios

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

zag

-0.77

áµ

-0.75

JV

-0.75

whatever

-0.71

ship

-0.71

addons

-0.70

Favorite

-0.70

species

-0.70

Í

-0.69

igmatic

-0.68

POSITIVE LOGITS

 asked

1.61

 confronted

1.51

 contacted

1.48

 questioned

1.47

 pressed

1.41

 approached

1.28

 inquired

1.18

 interviewed

1.18

 quizz

1.18

 challenged

1.14

Activations Density 0.109%

phrases indicating a change or event occurring in a sequence

instances of the word "when" used to indicate time or conditional scenarios

No Comments

No Known Activations