INDEX

Explanations

statements predicting or discussing potential future actions or consequences

oai_token-act-pair · gpt-3.5-turbo

words that indicate future actions or predictions

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

Plan

-0.68

engineering

-0.67

Reporting

-0.66

Methods

-0.66

eatured

-0.65

 Measures

-0.65

quished

-0.63

amped

-0.63

 Theme

-0.63

sembly

-0.62

POSITIVE LOGITS

 gladly

1.15

 surely

1.13

 inevitably

1.13

 undoubtedly

1.10

 doubtless

1.10

be

1.08

 likely

1.03

 probably

1.02

 eventually

1.02

 never

1.00

Activations Density 0.213%

statements predicting or discussing potential future actions or consequences

words that indicate future actions or predictions

No Comments

No Known Activations