INDEX

Explanations

evaluative language indicating opinions or judgments

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

0:0.01

1:0.32

2:0.08

3:0.03

4:0.01

5:0.07

6:0.09

7:0.04

8:0.08

9:0.10

10:0.07

11:0.04

Negative Logits

resy

-1.51

bats

-1.40

pps

-1.30

 harassed

-1.28

 Vampire

-1.28

 Pirate

-1.28

tests

-1.28

ritch

-1.26

achev

-1.22

erville

-1.22

POSITIVE LOGITS

ELY

1.48

GMT

1.47

*)

1.45

)}

1.42

 gram

1.39

IU

1.39

OE

1.34

�

1.34

)—

1.30

OPA

1.29

Activations Density 0.578%

evaluative language indicating opinions or judgments

No Comments

No Known Activations