INDEX

Explanations

phrases expressing a sense of inability or prohibition

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

0:0.03

1:0.02

2:0.10

3:0.23

4:0.02

5:0.05

6:0.05

7:0.05

8:0.06

9:0.08

10:0.14

11:0.11

Negative Logits

ullah

-1.03

 accrued

-0.97

heading

-0.94

trip

-0.92

rant

-0.91

BTC

-0.90

rued

-0.87

Pakistan

-0.86

 Laksh

-0.85

 hereafter

-0.84

POSITIVE LOGITS

CHA

1.10

 afford

1.06

�醒

1.02

 Oops

1.02

ties

1.01

ppo

0.99

keye

0.97

CHAT

0.96

agy

0.95

uate

0.95

Activations Density 0.027%

phrases expressing a sense of inability or prohibition

No Comments

No Known Activations