INDEX

Explanations

references to physical size or intensity

oai_token-act-pair · gpt-3.5-turbo

instances of the word "giant" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

yrinth

-0.92

iggins

-0.83

endment

-0.82

qi

-0.77

blance

-0.76

rences

-0.76

nc

-0.75

iring

-0.75

earchers

-0.74

tein

-0.73

POSITIVE LOGITS

 squid

0.93

 chunk

0.81

gest

0.78

 monster

0.78

 conglomer

0.77

 gorilla

0.77

 leap

0.75

 Slayer

0.72

 eyeb

0.72

 titan

0.71

Activations Density 0.017%

references to physical size or intensity

instances of the word "giant" in various contexts

No Comments

No Known Activations