INDEX

Explanations

names and entities, particularly those associated with a certain context or narrative

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_L2_Hcat_z_lr1.20e-03_l11.00e+00_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v4.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.2.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

0:0.08

1:0.03

2:0.25

3:0.05

4:0.13

5:0.07

6:0.03

7:0.03

8:0.04

9:0.17

10:0.04

11:0.02

Negative Logits

 newsp

-1.72

inventoryQuantity

-1.70

nces

-1.56

uations

-1.36

 horizont

-1.33

vez

-1.25

ioxide

-1.25

uyomi

-1.24

uously

-1.22

quickShipAvailable

-1.19

POSITIVE LOGITS

 Mara

1.42

Mesh

1.36

Tools

1.33

 Refuge

1.22

gat

1.21

 Peninsula

1.18

 Mamm

1.15

 Mech

1.13

pton

1.13

NSW

1.13

Activations Density 0.011%

names and entities, particularly those associated with a certain context or narrative

No Comments

No Known Activations