INDEX

Explanations

letters and numbers used in specific patterns

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

1:0.03

2:0.17

3:0.12

4:0.15

5:0.11

6:0.06

7:0.02

8:0.08

9:0.07

10:0.05

11:0.02

Negative Logits

zyk

-1.32

isers

-1.22

claw

-1.18

 Dresden

-1.13

hess

-1.09

 Brighton

-1.09

Khe

-1.07

Eid

-1.06

Yus

-1.05

Alz

-1.05

POSITIVE LOGITS

️

1.15

�

1.10

TT

1.09

à

1.07

illes

1.06

 Caption

1.06

HHHH

1.05

angled

1.04

rican

1.04

 denotes

1.04

Activations Density 0.009%

letters and numbers used in specific patterns

No Comments

No Known Activations