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

0:0.07

1:0.07

2:0.08

3:0.08

4:0.09

5:0.08

6:0.08

7:0.08

8:0.08

9:0.08

10:0.07

11:0.08

Negative Logits

jad

-3.02

Ka

-2.85

 princ

-2.75

SUN

-2.74

�

-2.73

clinton

-2.73

Moh

-2.70

Hur

-2.70

taboola

-2.65

 Hussein

-2.63

POSITIVE LOGITS

 Mech

2.93

PVC

2.84

controller

2.75

 Transformers

2.73

 Regulation

2.71

 mould

2.69

yna

2.55

 filament

2.48

 transformer

2.48

rider

2.48

Activations Density 0.000%

No Known Activations

This feature has no known activations.