INDEX

Explanations

mentions related to illegal activities, investigations, or scandals

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "extort" and its variations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

Kaw

-0.66

mar

-0.63

 dash

-0.63

pul

-0.62

 Rising

-0.61

dec

-0.60

 ledge

-0.60

sab

-0.60

Smy

-0.60

 chip

-0.59

POSITIVE LOGITS

orted

4.10

orting

3.84

orts

2.73

ort

2.40

orter

1.86

ORT

1.79

ortment

1.63

ortion

1.59

orters

1.46

ORTS

1.25

Activations Density 0.019%

mentions related to illegal activities, investigations, or scandals

occurrences of the word "extort" and its variations

No Comments

No Known Activations