INDEX

Explanations

negation or words indicating limits and exclusions

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_L3_Hcat_z_lr1.20e-03_l19.00e-01_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.3.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

0:0.02

1:0.06

2:0.22

3:0.13

4:0.01

5:0.02

6:0.07

7:0.06

8:0.14

9:0.09

10:0.07

11:0.06

Negative Logits

rique

-1.06

mar

-1.06

 Dominican

-0.92

ropolitan

-0.88

ocracy

-0.88

opal

-0.88

 Teddy

-0.86

Town

-0.86

vantage

-0.85

arus

-0.85

POSITIVE LOGITS

Reviewed

1.03

 behaved

1.00

vou

0.98

 sake

0.94

��

0.94

 grap

0.92

 mistaken

0.89

 サーティワン

0.89

 weren

0.88

 abund

0.87

Activations Density 0.042%

negation or words indicating limits and exclusions

No Comments

No Known Activations