INDEX

Explanations

phrases related to destruction or negative actions

oai_token-act-pair · gpt-3.5-turbo

instances of the word "destroy" 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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ETF

-0.88

 pleading

-0.76

culation

-0.75

annis

-0.74

legal

-0.73

isoft

-0.70

çīĪ

-0.69

cul

-0.66

ramid

-0.65

bid

-0.65

POSITIVE LOGITS

 havoc

0.90

roying

0.81

 destroy

0.81

 warheads

0.78

ãĥĥãĥī

0.73

 destroyer

0.70

ãĥĩãĤ£

0.69

 destroys

0.69

ishing

0.69

ishers

0.67

Activations Density 0.015%

phrases related to destruction or negative actions

instances of the word "destroy" in various contexts

No Comments

No Known Activations