INDEX

Explanations

references to specific names and titles

oai_token-act-pair · gpt-3.5-turbo

sequences of letters or characters indicative of specific phrases or names

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

raped

-0.68

 acute

-0.61

 Akin

-0.58

 specificity

-0.57

 respectively

-0.56

cause

-0.54

 afterlife

-0.54

 overboard

-0.52

 horr

-0.52

 shame

-0.52

POSITIVE LOGITS

phant

0.89

imaru

0.85

velt

0.81

tes

0.75

merce

0.74

Unix

0.72

nic

0.72

Redditor

0.72

SpaceEngineers

0.71

ois

0.70

Activations Density 0.353%

references to specific names and titles

sequences of letters or characters indicative of specific phrases or names

No Comments

No Known Activations