INDEX

Explanations

proper nouns, particularly names of characters, places, and titles in various contexts

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_L3_Hcat_z_lr1.20e-03_l19.00e-01_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.3.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

0:0.05

1:0.08

2:0.10

3:0.07

4:0.03

5:0.04

6:0.19

7:0.08

8:0.09

9:0.04

10:0.09

11:0.10

Negative Logits

 partition

-1.17

 redistributed

-1.08

ificant

-1.02

 unde

-1.01

 contiguous

-1.00

 varies

-0.96

 vary

-0.94

 varied

-0.94

anwhile

-0.94

 connectivity

-0.92

POSITIVE LOGITS

Tube

1.16

Jam

1.08

Strike

1.07

 Yards

1.07

 Hits

1.07

◼

1.07

 Velvet

1.06

Gate

1.06

kees

1.06

 Cups

1.06

Activations Density 0.494%

proper nouns, particularly names of characters, places, and titles in various contexts

No Comments

No Known Activations