INDEX

Explanations

phrases indicating action or significant events

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_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.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

1:0.01

2:0.21

3:0.14

4:0.14

5:0.04

6:0.07

7:0.08

8:0.04

9:0.07

10:0.08

11:0.04

Negative Logits

 Parables

-1.53

 Jinn

-1.51

 Amid

-1.44

 Representative

-1.36

Bid

-1.34

 Cait

-1.34

 fanc

-1.31

Tud

-1.30

 Fraz

-1.28

 qualifying

-1.28

POSITIVE LOGITS

ム

1.77

qqa

1.73

wx

1.68

']

1.60

oad

1.59

igon

1.55

keyes

1.54

rir

1.50

[/

1.50

ディ

1.48

Activations Density 0.001%

phrases indicating action or significant events

No Comments

No Known Activations