INDEX

Explanations

phrases indicating impact or consequences related to various subjects

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

0:0.08

1:0.09

2:0.02

3:0.13

4:0.08

5:0.11

6:0.04

7:0.02

8:0.27

9:0.06

10:0.02

11:0.02

Negative Logits

 repertoire

-2.07

 cass

-1.82

string

-1.80

asking

-1.79

prep

-1.73

perfect

-1.73

 promising

-1.72

imens

-1.71

aturdays

-1.68

pires

-1.68

POSITIVE LOGITS

 brunt

2.46

 adverse

2.42

 Factors

2.35

 injuries

2.05

 affected

2.04

 suffer

2.02

 Affect

2.02

 impacted

1.99

 absentee

1.97

effects

1.94

Activations Density 0.007%

phrases indicating impact or consequences related to various subjects

No Comments

No Known Activations