INDEX

Explanations

mentions of directors associated with films

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

Head Attr Weights

0:0.04

1:0.03

2:0.07

3:0.09

4:0.12

5:0.05

6:0.04

7:0.12

8:0.04

9:0.04

10:0.22

11:0.10

Negative Logits

raq

-1.87

iasis

-1.71

 graphene

-1.59

 phyl

-1.52

ratom

-1.51

 attest

-1.46

lux

-1.46

 libraries

-1.45

 Libraries

-1.45

 apples

-1.43

POSITIVE LOGITS

Direct

1.66

 Wichita

1.59

 Hitchcock

1.52

Jol

1.50

 Buckley

1.49

 Enrique

1.46

 Sunrise

1.45

 Dianne

1.43

 northwest

1.43

 staffed

1.43

Activations Density 0.003%

mentions of directors associated with films

No Known Activations