INDEX

Explanations

the infinitive form of verbs, specifically the word "to."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L9_Hcat_z_lr1.20e-03_l11.20e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.9.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

1:0.08

2:0.08

3:0.08

4:0.07

5:0.10

6:0.08

7:0.07

8:0.08

9:0.07

10:0.09

11:0.08

Negative Logits

@@

-2.67

 corrosion

-2.61

 sockets

-2.56

 steel

-2.54

osen

-2.52

strength

-2.51

rust

-2.51

 corro

-2.49

 inspecting

-2.44

 rust

-2.43

POSITIVE LOGITS

 Savage

2.93

ahu

2.74

Agg

2.73

FactoryReloaded

2.69

 monet

2.60

aez

2.58

Cub

2.55

eka

2.54

 Combo

2.50

 Mara

2.47

Activations Density 0.000%

the infinitive form of verbs, specifically the word "to."

No Comments

No Known Activations