INDEX

Explanations

phrases related to guiding the reader's attention or providing warnings

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.04

1:0.02

2:0.06

3:0.17

4:0.10

5:0.05

6:0.14

7:0.04

8:0.06

9:0.09

10:0.09

11:0.10

Negative Logits

 buckets

-1.55

TTL

-1.42

 Titanic

-1.40

luster

-1.36

 Genie

-1.29

 Bounty

-1.28

 garbage

-1.27

 bucket

-1.27

Boo

-1.26

 Bucket

-1.25

POSITIVE LOGITS

llor

1.44

cent

1.40

ghai

1.40

doi

1.33

]-

1.32

ndra

1.31

iven

1.31

np

1.31

ior

1.28

ivot

1.26

Activations Density 0.075%

phrases related to guiding the reader's attention or providing warnings

No Comments

No Known Activations