INDEX

Explanations

phrases that indicate a specific timeframe in the past

oai_token-act-pair · gpt-3.5-turbo

the article "a" in various contexts

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

Catal

-0.84

AIDS

-0.83

Edit

-0.83

culosis

-0.72

EU

-0.71

TPP

-0.71

independence

-0.70

ahime

-0.70

achine

-0.69

evidence

-0.69

POSITIVE LOGITS

 handful

1.13

 bunch

1.08

lot

0.99

 fraction

0.99

tad

0.98

few

0.97

 couple

0.96

bit

0.90

 temporary

0.89

 dozen

0.89

Activations Density 0.130%

phrases that indicate a specific timeframe in the past

the article "a" in various contexts

No Comments

No Known Activations