INDEX

Explanations

phrases that indicate actions or measures that have been taken or need to be taken

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

0:0.35

1:0.01

2:0.17

3:0.09

4:0.03

5:0.04

6:0.01

7:0.04

8:0.02

9:0.01

10:0.14

11:0.02

Negative Logits

Winner

-2.69

 Writers

-2.65

Writ

-2.64

 Fiction

-2.64

 cartoon

-2.51

 columnist

-2.49

 slot

-2.46

 pund

-2.44

�

-2.41

Column

-2.41

POSITIVE LOGITS

 steps

3.92

 Steps

3.86

 precaution

3.25

autions

3.10

 precautions

3.09

 undertaken

3.06

 STEP

2.95

 Step

2.91

 mitigation

2.88

steps

2.83

Activations Density 0.018%

phrases that indicate actions or measures that have been taken or need to be taken

No Comments

No Known Activations