INDEX

Explanations

sequences indicating political or controversial discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L10_Hcat_z_lr1.20e-03_l11.30e+00_ds24576_bs4096_dc1.00e-05_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.10.attn.hook_z

Hook Layer

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.32

1:0.07

2:0.09

3:0.06

4:0.04

5:0.02

6:0.06

7:0.07

8:0.02

9:0.05

10:0.10

11:0.04

Negative Logits

 Floyd

-4.74

 Mayweather

-3.90

Indian

-3.67

 OnePlus

-3.61

 Zimbabwe

-3.58

 Quan

-3.50

 Cherokee

-3.44

ucha

-3.42

PW

-3.36

Mex

-3.35

POSITIVE LOGITS

 Anders

9.37

å

8.01

 Sven

7.90

 Stockholm

7.76

 Swed

7.57

 Swedish

7.49

Bj

7.40

 Sweden

7.30

 Norwegian

6.98

 Oslo

6.87

Activations Density 0.166%

sequences indicating political or controversial discussions

No Comments

No Known Activations