INDEX

Explanations

words and phrases related to political affiliations and opinions

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_L3_Hcat_z_lr1.20e-03_l19.00e-01_ds24576_bs4096_dc1.00e-06_rsanthropic_rie25000_nr4_v9.pt

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

float32

Hook Name

blocks.3.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

0:0.01

1:0.05

2:0.17

3:0.05

4:0.02

5:0.07

6:0.06

7:0.06

8:0.16

9:0.11

10:0.07

11:0.10

Negative Logits

''.

-1.41

 earthqu

-1.29

availability

-1.29

 unden

-1.07

 secondly

-1.06

umbered

-1.06

eway

-1.05

sol

-1.05

 equival

-1.05

Required

-1.05

POSITIVE LOGITS

unts

1.31

governmental

1.15

otics

1.14

BUR

1.12

burgh

1.11

ulz

1.08

aser

1.08

rica

1.06

asers

1.06

rams

1.05

Activations Density 0.050%

words and phrases related to political affiliations and opinions

No Comments

No Known Activations