INDEX

Explanations

references to categories or types of things

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

2:0.15

3:0.21

4:0.02

5:0.02

6:0.14

7:0.08

8:0.04

9:0.03

10:0.09

11:0.11

Negative Logits

uel

-1.26

eka

-1.23

rahim

-1.14

ique

-1.13

 outweigh

-1.09

arcity

-1.08

nih

-1.07

neys

-1.05

rolet

-1.03

dyl

-1.01

POSITIVE LOGITS

 horm

1.05

 resumes

0.97

 shenan

0.96

 secretion

0.96

 honors

0.95

 Izan

0.93

 shenanigans

0.93

 imaginable

0.93

 stimuli

0.88

 horr

0.87

Activations Density 0.006%

references to categories or types of things

No Comments

No Known Activations