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

1:0.11

2:0.07

3:0.08

4:0.08

5:0.07

6:0.08

7:0.08

8:0.07

9:0.06

10:0.09

11:0.08

Negative Logits

continental

-1.67

 cannabin

-1.61

 indic

-1.60

azines

-1.58

 retina

-1.57

arden

-1.57

 textile

-1.55

endon

-1.51

 mater

-1.48

 singles

-1.45

POSITIVE LOGITS

gaard

1.88

STEP

1.80

HAHAHAHA

1.62

sung

1.62

WAY

1.61

APTER

1.59

zees

1.58

phabet

1.58

 "$:/

1.58

chall

1.57

Activations Density 0.000%

No Known Activations

This feature has no known activations.