INDEX

Explanations

the word "so" occurring with high activation values

oai_token-act-pair · gpt-3.5-turbo

phrases that express a sense of negation or contradiction

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

ãĥ¼ãĥĨãĤ£

-0.85

é¾

-0.69

MAP

-0.68

ãĤ¼ãĤ¦ãĤ¹

-0.64

 annotations

-0.63

ayne

-0.60

{:

-0.57

 SHARES

-0.57

DERR

-0.57

 scan

-0.56

POSITIVE LOGITS

 much

0.86

aked

0.84

oths

0.82

oooo

0.81

ppy

0.80

othes

0.77

zin

0.76

akers

0.74

icably

0.73

oooooooo

0.73

Activations Density 0.061%

the word "so" occurring with high activation values

phrases that express a sense of negation or contradiction

No Comments

No Known Activations