INDEX

Explanations

positive descriptors and evaluations related to people and their actions

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

0:0.02

1:0.07

2:0.23

3:0.04

4:0.02

5:0.03

6:0.09

7:0.08

8:0.11

9:0.09

10:0.08

11:0.08

Negative Logits

 settlements

-1.24

aughtered

-1.22

apons

-1.12

anches

-1.10

 corridors

-1.10

ockets

-1.09

onies

-1.09

Oak

-1.06

abases

-1.06

 horns

-1.04

POSITIVE LOGITS

Reviewer

1.25

 fuss

1.15

 Psychiat

1.12

#$

1.11

 Perspective

1.11

!?

1.04

 Sandwich

1.03

�

1.02

 compliment

1.02

 fool

1.02

Activations Density 0.022%

positive descriptors and evaluations related to people and their actions

No Comments

No Known Activations