INDEX

Explanations

information related to steps or instructions

oai_token-act-pair · gpt-3.5-turbo

phrases related to rules and limits

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

umni

-0.65

inis

-0.59

yssey

-0.55

ERA

-0.53

VERTISEMENT

-0.53

Ire

-0.53

tips

-0.52

sequently

-0.52

ramid

-0.52

Fre

-0.51

POSITIVE LOGITS

 oneself

0.61

 finite

0.60

 shitty

0.58

 crappy

0.57

 wrong

0.55

 objectively

0.55

 boring

0.54

 rationality

0.53

 systematically

0.53

 perceptual

0.52

Activations Density 2.249%

information related to steps or instructions

phrases related to rules and limits

No Comments

No Known Activations