INDEX

Explanations

verbs related to complete removal or destruction, often with negative connotations

oai_token-act-pair · gpt-3.5-turbo

references to the action of removing or erasing something

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

anooga

-1.03

egal

-0.80

 skelet

-0.80

annis

-0.69

cmp

-0.67

ideos

-0.66

pheus

-0.66

ensical

-0.65

trak

-0.65

ansas

-0.63

POSITIVE LOGITS

 wipe

1.14

 wiping

1.08

 wipes

1.03

 wiped

0.94

 erase

0.84

uto

0.77

ipe

0.77

 goodbye

0.76

::::::::

0.76

undo

0.72

Activations Density 0.006%

verbs related to complete removal or destruction, often with negative connotations

references to the action of removing or erasing something

No Comments

No Known Activations