INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_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.9.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

0:0.08

1:0.08

2:0.08

3:0.08

4:0.08

5:0.09

6:0.09

7:0.08

8:0.07

9:0.08

10:0.06

11:0.08

Negative Logits

 clutch

-2.75

 Scythe

-2.64

 reliability

-2.56

 Vict

-2.38

 Crash

-2.38

 reader

-2.35

 appl

-2.34

 brute

-2.34

 Souls

-2.33

(&

-2.31

POSITIVE LOGITS

aghd

3.23

eneg

2.93

issions

2.78

izophren

2.78

�

2.73

rients

2.72

=-=-=-=-

2.65

loe

2.65

 Mongol

2.64

separ

2.59

Activations Density 0.000%

No Known Activations

This feature has no known activations.