INDEX

Explanations

phrases related to editing or modifying content

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

1:0.03

2:0.28

3:0.04

4:0.01

5:0.04

6:0.09

7:0.05

8:0.14

9:0.11

10:0.08

11:0.06

Negative Logits

 Bounty

-1.34

Wat

-1.31

 Boat

-1.20

Bug

-1.18

 Fell

-1.18

Bug

-1.11

 naming

-1.08

Jur

-1.05

 Kamp

-1.03

Mer

-1.03

POSITIVE LOGITS

 safely

1.76

 cheaply

1.61

 efficiently

1.53

ociated

1.44

 easily

1.39

ulated

1.34

ilaterally

1.32

quished

1.31

quist

1.31

ades

1.29

Activations Density 0.117%

phrases related to editing or modifying content

No Comments

No Known Activations