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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.10

1:0.10

2:0.07

3:0.10

4:0.07

5:0.08

6:0.07

7:0.05

8:0.09

9:0.08

10:0.06

11:0.08

Negative Logits

Roaming

-1.97

 hers

-1.81

FN

-1.74

 Flavoring

-1.65

�

-1.61

 Blanc

-1.55

Grand

-1.55

Recommend

-1.52

 Torch

-1.46

ADRA

-1.44

POSITIVE LOGITS

gaard

1.96

adobe

1.82

erion

1.69

uph

1.67

ali

1.66

 hurd

1.63

eas

1.62

ascus

1.58

ctors

1.58

hod

1.55

Activations Density 0.000%

No Known Activations

This feature has no known activations.