INDEX

Explanations

names of individuals and their associated statements

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_L9_Hcat_z_lr1.20e-03_l11.20e+00_ds24576_bs4096_dc1.00e-06_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.9.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

0:0.13

1:0.02

2:0.06

3:0.18

4:0.06

5:0.03

6:0.01

7:0.04

8:0.02

9:0.01

10:0.37

11:0.01

Negative Logits

 Associates

-2.15

profits

-2.15

 commissions

-2.09

 collectors

-2.01

 committees

-2.01

 supervisors

-1.92

 duplication

-1.91

 colle

-1.91

wcsstore

-1.91

 laboratories

-1.90

POSITIVE LOGITS

 joked

2.68

 uttered

2.65

 chanted

2.60

 told

2.50

 muttered

2.45

 confidently

2.45

 said

2.42

iannopoulos

2.39

 shouted

2.38

 vowed

2.37

Activations Density 0.065%

names of individuals and their associated statements

No Comments

No Known Activations