INDEX

Explanations

phrases related to rules, regulations, and organizational processes

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.02

1:0.01

2:0.08

3:0.07

4:0.24

5:0.03

6:0.03

7:0.28

8:0.03

9:0.04

10:0.07

11:0.05

Negative Logits

rella

-1.78

amon

-1.72

pired

-1.72

riter

-1.70

arta

-1.60

orses

-1.58

reon

-1.57

verbs

-1.57

Adventure

-1.54

famous

-1.51

POSITIVE LOGITS

 fatigue

1.82

 imbalance

1.71

 troublesome

1.62

 toxicity

1.61

 defect

1.60

 deficiencies

1.60

 rash

1.59

 wrongdoing

1.57

 offending

1.54

 withdrawals

1.50

Activations Density 0.001%

phrases related to rules, regulations, and organizational processes

No Comments

No Known Activations