INDEX

Explanations

text related to classified information

oai_token-act-pair · gpt-3.5-turbo

references to classified or sensitive information

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

Wave

-0.68

LV

-0.67

 Stamp

-0.67

vous

-0.65

 Bale

-0.63

 Ches

-0.63

vas

-0.62

 Farmer

-0.62

 Atkinson

-0.61

Pause

-0.61

POSITIVE LOGITS

 classified

1.17

classified

1.06

 classify

0.95

 briefings

0.94

 warr

0.84

 veter

0.82

 newsp

0.82

iaries

0.81

 markings

0.79

 classification

0.78

Activations Density 0.007%

text related to classified information

references to classified or sensitive information

No Comments

No Known Activations