INDEX

Explanations

names or terms related to people or medical conditions

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_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

0:0.06

1:0.04

2:0.23

3:0.07

4:0.17

5:0.05

6:0.02

7:0.04

8:0.10

9:0.10

10:0.05

11:0.03

Negative Logits

Ago

-1.26

interstitial

-1.25

imar

-1.21

hement

-1.18

iard

-1.17

quartered

-1.15

 bishops

-1.14

��

-1.14

 Churches

-1.10

nai

-1.09

POSITIVE LOGITS

insk

1.22

QL

1.18

Balt

1.16

────

1.16

Spec

1.13

ormon

1.11

osis

1.09

Character

1.08

Characters

1.07

Ranked

1.07

Activations Density 0.001%

names or terms related to people or medical conditions

No Comments

No Known Activations