INDEX

Explanations

phrases related to staying away or avoiding something

oai_token-act-pair · gpt-3.5-turbo

phrases related to avoidance or staying away from something

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

catentry

-0.78

elf

-0.74

 roundup

-0.72

lot

-0.67

 Navajo

-0.62

MS

-0.62

erity

-0.60

 Surprise

-0.59

 converter

-0.58

 profiling

-0.57

POSITIVE LOGITS

 indefinitely

0.85

lishes

0.82

 confines

0.78

andestine

0.77

 bounds

0.74

 forever

0.73

 haun

0.69

angered

0.68

arcer

0.67

 intact

0.67

Activations Density 0.134%

phrases related to staying away or avoiding something

phrases related to avoidance or staying away from something

No Comments

No Known Activations