INDEX

Explanations

the word "to."

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

0:0.09

1:0.06

2:0.08

3:0.09

4:0.08

5:0.08

6:0.08

7:0.07

8:0.08

9:0.07

10:0.10

11:0.08

Negative Logits

 cling

-1.63

rette

-1.55

CAD

-1.53

 itch

-1.51

 Edition

-1.48

 defective

-1.45

Downloadha

-1.44

 Shutterstock

-1.44

 matt

-1.43

ATM

-1.43

POSITIVE LOGITS

 Leilan

1.83

Split

1.79

Bye

1.75

leigh

1.73

iste

1.68

�

1.66

Dialogue

1.62

Hunter

1.58

govtrack

1.58

EVA

1.57

Activations Density 0.000%

the word "to."

No Comments

No Known Activations