INDEX

Explanations

emphasizing words indicating certainty or quality assessments

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

IFrame

Link

Not in Any Lists

No Comments

Head Attr Weights

0:0.22

1:0.04

2:0.03

3:0.08

4:0.07

5:0.05

6:0.05

7:0.08

8:0.06

9:0.21

10:0.03

11:0.03

Negative Logits

olphins

-2.15

affle

-1.91

ruce

-1.88

yssey

-1.87

aths

-1.73

ittee

-1.73

ollower

-1.72

ricanes

-1.71

illon

-1.69

opus

-1.69

POSITIVE LOGITS

 acceptable

1.94

 nonetheless

1.91

 habitable

1.84

 worthy

1.81

 enough

1.79

worthy

1.78

huh

1.75

 Stam

1.75

acceptable

1.74

worldly

1.74

Activations Density 0.007%

emphasizing words indicating certainty or quality assessments

No Comments

No Known Activations