INDEX

Explanations

associated with physical barriers or obstacles

oai_token-act-pair · gpt-3.5-turbo

references to obstacles or hindrances

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ivery

-0.84

iability

-0.79

phis

-0.76

sch

-0.72

inav

-0.70

ergy

-0.69

orp

-0.69

sb

-0.68

ovie

-0.68

ribution

-0.68

POSITIVE LOGITS

 barriers

1.11

 barrier

0.98

 erected

0.96

 separating

0.88

 walls

0.87

 Walls

0.86

 Barrier

0.81

buster

0.79

 wall

0.78

 fence

0.73

Activations Density 0.016%

associated with physical barriers or obstacles

references to obstacles or hindrances

No Comments

No Known Activations