INDEX

Explanations

instances of the special character "âĢ" 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_L5_Hcat_z_lr1.20e-03_l11.00e+00_ds49152_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

49,152

Data Type

float32

Hook Name

blocks.5.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

0:0.07

1:0.07

2:0.06

3:0.05

4:0.07

5:0.13

6:0.37

7:0.02

8:0.05

9:0.02

10:0.03

11:0.02

Negative Logits

 interference

-1.56

 placement

-1.40

 gaps

-1.39

geries

-1.35

 shortages

-1.33

 saturation

-1.32

 complexity

-1.31

 similarities

-1.31

 discrepancies

-1.30

 specials

-1.28

POSITIVE LOGITS

agine

1.98

arest

1.77

wu

1.75

 admire

1.62

ynes

1.61

efer

1.55

cients

1.53

ould

1.47

�

1.47

ioned

1.46

Activations Density 0.018%

instances of the special character "âĢ" in various contexts

No Comments

No Known Activations