INDEX

Explanations

negations and qualifiers that express low certainty or limitations

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

1:0.08

2:0.13

3:0.03

4:0.02

5:0.04

6:0.16

7:0.06

8:0.18

9:0.12

10:0.06

11:0.04

Negative Logits

ameron

-1.15

ials

-1.08

alks

-1.02

 bush

-1.01

 endings

-0.99

actions

-0.99

raz

-0.98

rez

-0.97

ival

-0.96

ATIONS

-0.95

POSITIVE LOGITS

orthy

1.16

 otherwise

1.15

pires

1.14

 aest

1.08

entimes

1.08

 geographically

1.08

 wearer

1.03

pport

1.03

 Depend

1.03

pire

1.01

Activations Density 0.210%

negations and qualifiers that express low certainty or limitations

No Comments

No Known Activations