INDEX

Explanations

elements related to authority or power dynamics

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

0:0.50

1:0.06

2:0.02

3:0.04

4:0.02

5:0.10

6:0.05

7:0.01

8:0.07

9:0.04

10:0.02

11:0.02

Negative Logits

 automation

-1.67

 rubble

-1.66

 PubMed

-1.63

 deflation

-1.61

 salvage

-1.61

 Reuters

-1.59

 oxygen

-1.58

 Codec

-1.54

 fodder

-1.53

jog

-1.53

POSITIVE LOGITS

agree

2.40

shall

2.23

ende

2.18

ree

2.13

ivan

2.12

want

2.11

being

2.11

she

2.10

few

2.07

ever

2.06

Activations Density 0.078%

elements related to authority or power dynamics

No Comments

No Known Activations