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

0:0.06

1:0.05

2:0.08

3:0.08

4:0.09

5:0.08

6:0.08

7:0.10

8:0.08

9:0.08

10:0.09

11:0.08

Negative Logits

FO

-2.03

FB

-1.86

��

-1.84

KING

-1.82

IGHTS

-1.80

aughters

-1.76

OTOS

-1.76

lli

-1.74

tv

-1.74

��

-1.71

POSITIVE LOGITS

 Notting

2.04

 Hazard

1.86

geries

1.75

 Pompe

1.70

 Burgess

1.70

Gau

1.69

 Peel

1.69

 Peck

1.67

 Bolton

1.67

 Prob

1.67

Activations Density 0.000%

No Known Activations

This feature has no known activations.