INDEX

Explanations

mentions of interviews

oai_token-act-pair · gpt-3.5-turbo

repeated references to interviews

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

axy

-0.82

cil

-0.80

¶ħ

-0.76

mil

-0.74

math

-0.74

hover

-0.70

ple

-0.68

ignt

-0.66

fal

-0.66

warm

-0.65

POSITIVE LOGITS

 interviews

1.06

 interview

1.00

 Interview

0.93

ees

0.92

Interview

0.87

ioned

0.86

 interviewing

0.80

 Transcript

0.78

 transcripts

0.76

annel

0.75

Activations Density 0.020%

mentions of interviews

repeated references to interviews

No Comments

No Known Activations