INDEX

Explanations

text related to written transcripts

oai_token-act-pair · gpt-3.5-turbo

mentions of transcripts

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

zzle

-0.69

vil

-0.69

Geh

-0.67

ÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂłÂł

-0.66

dream

-0.66

inence

-0.65

yu

-0.64

pton

-0.64

 Slug

-0.63

 Prevention

-0.63

POSITIVE LOGITS

 transcripts

1.28

 Transcript

1.20

 transcript

1.18

 transcription

1.04

ions

0.86

 tapes

0.85

 snippets

0.85

ophon

0.84

 recordings

0.84

ophone

0.82

Activations Density 0.020%

text related to written transcripts

mentions of transcripts

No Comments

No Known Activations