INDEX

Explanations

phrases related to making mistakes or causing trouble

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

vation

-0.91

Interstitial

-0.85

rica

-0.69

uti

-0.64

ãĥ´

-0.63

 Citation

-0.62

Ļ

-0.61

eele

-0.60

Ĺ

-0.58

oko

-0.58

POSITIVE LOGITS

around

0.98

 around

0.97

 havoc

0.97

 Around

0.85

driver

0.77

bley

0.75

Around

0.75

up

0.71

ily

0.71

ishly

0.70

Activations Density 0.081%

phrases related to making mistakes or causing trouble

No Comments

No Known Activations