INDEX

Explanations

the word "wipe" or its variations

oai_token-act-pair · gpt-3.5-turbo

actions associated with erasure or elimination

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

fman

-0.83

pheus

-0.72

udeau

-0.71

iterranean

-0.71

FML

-0.70

ensical

-0.70

anooga

-0.69

 guiActiveUnfocused

-0.67

ional

-0.64

egal

-0.64

POSITIVE LOGITS

 clean

0.90

 away

0.86

 wiped

0.86

awa

0.85

 wiping

0.84

 wipe

0.83

 havoc

0.83

 wipes

0.82

 ashore

0.82

 slate

0.82

Activations Density 0.030%

the word "wipe" or its variations

actions associated with erasure or elimination

No Comments

No Known Activations