INDEX

Explanations

phrases that indicate prediction or speculation about future events

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_L9_Hcat_z_lr1.20e-03_l11.20e+00_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.9.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

1:0.02

2:0.09

3:0.28

4:0.02

5:0.05

6:0.02

7:0.05

8:0.04

9:0.01

10:0.30

11:0.01

Negative Logits

 Crimes

-2.19

OWN

-2.00

ankind

-1.97

],

-1.96

 mankind

-1.92

 Killed

-1.92

Destroy

-1.90

Technology

-1.89

 campuses

-1.87

 Destroy

-1.87

POSITIVE LOGITS

 guess

2.73

 accordingly

2.70

 definitely

2.69

 guesses

2.66

 wisely

2.62

 luck

2.60

 guessing

2.54

 nicely

2.51

 awfully

2.50

 gotta

2.45

Activations Density 0.272%

phrases that indicate prediction or speculation about future events

No Comments

No Known Activations