INDEX

Explanations

issues related to legal or procedural accountability

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.03

1:0.02

2:0.30

3:0.15

4:0.12

5:0.02

6:0.03

7:0.08

8:0.07

9:0.02

10:0.04

11:0.06

Negative Logits

rote

-2.02

renches

-1.90

lite

-1.86

igate

-1.84

ighters

-1.77

rehensive

-1.75

��

-1.70

ILLE

-1.69

ighter

-1.66

eper

-1.65

POSITIVE LOGITS

 constraints

1.72

 Scand

1.72

 penchant

1.70

 propensity

1.63

 uncertainties

1.49

 sterling

1.48

imus

1.47

 circumstances

1.46

 looming

1.44

 realities

1.43

Activations Density 0.013%

issues related to legal or procedural accountability

No Comments

No Known Activations