INDEX

Explanations

phrases indicating clarification or emphasis on specific details

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.03

1:0.03

2:0.06

3:0.19

4:0.09

5:0.08

6:0.05

7:0.03

8:0.07

9:0.14

10:0.11

11:0.07

Negative Logits

[|

-1.61

FTWARE

-1.58

ali

-1.49

*.

-1.46

 fidelity

-1.28

 [...]

-1.26

 emulation

-1.26

 […]

-1.26

 independ

-1.24

pc

-1.21

POSITIVE LOGITS

 Bridges

1.57

swers

1.50

 reacted

1.33

pees

1.32

 Brock

1.32

 laughed

1.30

 countered

1.30

 greeted

1.28

?]

1.28

 jurors

1.27

Activations Density 0.295%

phrases indicating clarification or emphasis on specific details

No Comments

No Known Activations