INDEX

Explanations

phrases that indicate time periods or dates

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

0:0.01

1:0.02

2:0.06

3:0.27

4:0.01

5:0.03

6:0.09

7:0.15

8:0.05

9:0.04

10:0.09

11:0.12

Negative Logits

ロ

-1.23

エル

-1.17

enthal

-1.15

QC

-1.12

chwitz

-1.10

 Enchant

-1.10

�

-1.05

 Cyborg

-1.05

ucci

-1.00

olini

-1.00

POSITIVE LOGITS

displayText

1.24

IPM

1.23

spection

1.06

 Macy

1.05

finals

1.03

ftime

1.03

torn

0.98

imester

0.97

 deliberations

0.97

Loading

0.96

Activations Density 0.002%

phrases that indicate time periods or dates

No Comments

No Known Activations