INDEX

Explanations

mentions or references to federal institutions or government-related activities

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Configuration

ckkissane/attn-saes-gpt2-small-all-layers/gpt2-small_L0_Hcat_z_lr1.20e-03_l11.80e+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.0.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

0:0.03

1:0.19

2:0.06

3:0.13

4:0.10

5:0.14

6:0.03

7:0.05

8:0.05

9:0.06

10:0.08

11:0.03

Negative Logits

lihood

-4.68

=-=-

-4.46

-4.15

hra

-3.89

Else

-3.84

points

-3.78

...]

-3.77

}"

-3.75

 Flavoring

-3.70

Stage

-3.56

POSITIVE LOGITS

 unemploy

4.32

 councill

4.29

ailability

4.15

isphere

4.12

 confir

3.98

 antidepress

3.85

 citiz

3.82

iazep

3.80

 undermin

3.79

 helicop

3.71

Activations Density 0.010%

mentions or references to federal institutions or government-related activities

No Comments

No Known Activations