INDEX

Explanations

names of people or characters in various contexts

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

0:0.29

1:0.05

2:0.05

3:0.06

4:0.04

5:0.08

6:0.04

7:0.03

8:0.06

9:0.11

10:0.10

11:0.04

Negative Logits

 geopolitical

-1.99

 ancest

-1.84

 transmit

-1.83

 euphem

-1.81

 transmitting

-1.80

 electoral

-1.78

 abbre

-1.76

 stimulus

-1.74

 mandate

-1.73

 acronym

-1.70

POSITIVE LOGITS

riel

2.34

ador

2.20

chel

2.16

loe

2.12

ora

2.11

agh

2.09

nce

2.06

ilda

2.05

cia

2.05

aea

2.03

Activations Density 0.305%

names of people or characters in various contexts

No Comments

No Known Activations