INDEX

Explanations

occurrences of the word "to" in various contexts

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

Head Attr Weights

0:0.03

1:0.01

2:0.11

3:0.04

4:0.07

5:0.03

6:0.06

7:0.13

8:0.04

9:0.03

10:0.09

11:0.30

Negative Logits

 vantage

-1.85

 bearings

-1.80

antage

-1.79

selage

-1.71

ront

-1.70

icago

-1.70

antry

-1.70

ocaly

-1.69

perial

-1.66

hent

-1.66

POSITIVE LOGITS

 Ronnie

1.75

 immunity

1.59

 affirmative

1.59

chance

1.58

 Petition

1.57

DEF

1.56

 lottery

1.56

 Derrick

1.54

 losers

1.51

 Immunity

1.50

Activations Density 0.003%

occurrences of the word "to" in various contexts

No Known Activations