INDEX

Explanations

proper nouns and phrases related to head-to-head comparisons

oai_token-act-pair · gpt-3.5-turbo

references to television shows and their elements

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

Riv

-0.82

PLA

-0.80

 FANT

-0.77

vette

-0.77

iev

-0.72

 Prot

-0.69

Els

-0.68

Dex

-0.67

VG

-0.66

Et

-0.65

POSITIVE LOGITS

 Head

2.08

Head

2.04

 Heads

2.04

 head

1.97

head

1.93

 HEAD

1.84

 heads

1.84

heads

1.73

HEAD

1.72

 tails

1.60

Activations Density 0.125%

proper nouns and phrases related to head-to-head comparisons

references to television shows and their elements

No Comments

No Known Activations