INDEX

Explanations

the word "go."

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_L7_Hcat_z_lr1.20e-03_l11.10e+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.7.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

0:0.08

1:0.07

2:0.06

3:0.09

4:0.07

5:0.08

6:0.08

7:0.07

8:0.09

9:0.09

10:0.08

11:0.07

Negative Logits

agall

-2.30

ギ

-2.15

encer

-2.10

 Leica

-2.06

legram

-1.98

ecast

-1.97

agen

-1.96

brush

-1.96

clipse

-1.94

ワン

-1.94

POSITIVE LOGITS

 decay

2.14

tha

2.02

,...

1.99

 winters

1.99

 commonly

1.93

 measurable

1.86

 aging

1.85

 nonetheless

1.83

 spikes

1.81

 periods

1.81

Activations Density 0.000%

the word "go."

No Comments

No Known Activations