INDEX

Explanations

content related to distorted information or views

oai_token-act-pair · gpt-3.5-turbo

terms related to distortion and its variations in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

ept

-0.78

git

-0.77

dain

-0.74

minent

-0.72

hal

-0.68

tained

-0.68

Flu

-0.67

ilet

-0.67

ille

-0.66

tes

-0.66

POSITIVE LOGITS

 distort

1.43

 distortion

1.42

 distortions

1.40

 distorted

1.25

 misunder

0.84

 havoc

0.84

 inaccur

0.80

 amplify

0.76

 exagger

0.74

 mislead

0.72

Activations Density 0.012%

content related to distorted information or views

terms related to distortion and its variations in various contexts

No Comments

No Known Activations