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

0:0.08

1:0.05

2:0.08

3:0.08

4:0.07

5:0.07

6:0.09

7:0.08

8:0.09

9:0.08

10:0.08

11:0.09

Negative Logits

��

-1.85

ocument

-1.64

[];

-1.52

hidden

-1.52

 Eliot

-1.51

":-

-1.49

iece

-1.46

 Strikes

-1.46

soever

-1.45

insk

-1.45

POSITIVE LOGITS

acid

1.69

 certification

1.68

cpu

1.66

bid

1.55

 wrist

1.54

af

1.53

 Compliance

1.52

 patent

1.49

 ammonia

1.49

bledon

1.49

Activations Density 0.000%

No Known Activations

This feature has no known activations.