INDEX

Explanations

references to organized groups or committees

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_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

0:0.03

1:0.02

2:0.10

3:0.07

4:0.06

5:0.04

6:0.06

7:0.30

8:0.03

9:0.05

10:0.13

11:0.05

Negative Logits

 reuse

-1.85

 gratification

-1.80

 mileage

-1.73

ocide

-1.65

nown

-1.59

 disrespect

-1.59

 unintended

-1.59

ocry

-1.53

 Wear

-1.53

 Practices

-1.51

POSITIVE LOGITS

 deliberations

1.85

soDeliveryDate

1.77

 finalists

1.75

 powerhouse

1.74

sembly

1.67

 comprising

1.66

 committee

1.64

 overseeing

1.63

atan

1.62

 benches

1.61

Activations Density 0.003%

references to organized groups or committees

No Comments

No Known Activations