INDEX

Explanations

phrases suggesting different options or choices for consideration

oai_token-act-pair · gpt-3.5-turbo

phrases that pose questions or seek engagement

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

odder

-0.65

Sense

-0.61

nih

-0.61

 Loaded

-0.60

Saw

-0.59

seller

-0.59

original

-0.59

cc

-0.58

 cannot

-0.57

aviour

-0.57

POSITIVE LOGITS

 congratulations

0.74

EStream

0.74

 congr

0.73

classes

0.70

 nomine

0.65

akeru

0.63

STEM

0.61

 peac

0.61

dinand

0.61

ð

0.60

Activations Density 0.023%

phrases suggesting different options or choices for consideration

phrases that pose questions or seek engagement

No Comments

No Known Activations