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

0:0.04

1:0.03

2:0.05

3:0.07

4:0.05

5:0.14

6:0.04

7:0.05

8:0.16

9:0.09

10:0.17

11:0.08

Negative Logits

 fragmentation

-1.78

 explosion

-1.75

 dominated

-1.73

alia

-1.73

 profiles

-1.72

 atro

-1.70

 infiltrated

-1.69

isphere

-1.69

 aggress

-1.69

 stagn

-1.67

POSITIVE LOGITS

etts

2.08

onne

1.96

 Restaur

1.90

 yarn

1.85

ormons

1.84

uitive

1.84

�

1.82

enegger

1.75

ithmetic

1.75

orah

1.72

Activations Density 0.000%

No Known Activations

This feature has no known activations.