INDEX

Explanations

No Explanations Found

New Auto-Interp

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

0:0.08

1:0.08

2:0.09

3:0.08

4:0.09

5:0.06

6:0.08

7:0.08

8:0.08

9:0.07

10:0.08

11:0.08

Negative Logits

phas

-1.74

 accordingly

-1.69

ement

-1.61

Interface

-1.54

ements

-1.53

Disable

-1.50

Effective

-1.48

versions

-1.47

 weaker

-1.45

Physical

-1.44

POSITIVE LOGITS

 Aires

2.02

ÃÂ

1.72

borough

1.71

inia

1.64

htaking

1.61

oops

1.61

 Investigative

1.59

ksh

1.57

ivan

1.56

 photographer

1.56

Activations Density 0.000%

No Known Activations

This feature has no known activations.