INDEX

Explanations

No Explanations Found

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.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

1:0.05

2:0.08

3:0.08

4:0.09

5:0.08

6:0.07

7:0.09

8:0.09

9:0.08

10:0.09

11:0.09

Negative Logits

 grooming

-1.61

aldehyde

-1.54

 council

-1.52

 bondage

-1.50

 councils

-1.49

ooters

-1.44

 destro

-1.42

HAEL

-1.40

Council

-1.39

 Corinth

-1.39

POSITIVE LOGITS

EStream

1.72

racuse

1.71

 guiActiveUn

1.63

 impl

1.62

Via

1.58

embed

1.55

 predicted

1.55

 playbook

1.52

lat

1.49

optim

1.46

Activations Density 0.000%

No Known Activations

This feature has no known activations.