INDEX

Explanations

references to flaunting or showcasing personal identity or attributes

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

New Auto-Interp

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.01

1:0.01

2:0.05

3:0.07

4:0.14

5:0.02

6:0.04

7:0.41

8:0.03

9:0.04

10:0.09

11:0.04

Negative Logits

 runoff

-1.89

vette

-1.59

imo

-1.56

jo

-1.53

arnaev

-1.52

osta

-1.44

Streamer

-1.41

 Stevens

-1.41

ero

-1.40

enf

-1.39

POSITIVE LOGITS

 secrecy

2.14

 unbeliev

1.69

 cumbers

1.63

 deception

1.62

interstitial

1.62

 behavi

1.61

 deceptive

1.59

 borders

1.58

 superiority

1.57

 Masquerade

1.56

Activations Density 0.001%

references to flaunting or showcasing personal identity or attributes

No Comments

No Known Activations