INDEX

Explanations

hypothetical situations or regrets in which actions could have been different

oai_token-act-pair · gpt-3.5-turbo

conditional statements and hypothetical scenarios involving outcomes

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

phies

-0.71

verend

-0.65

cial

-0.63

 nightly

-0.62

orate

-0.62

orship

-0.61

nex

-0.59

enos

-0.57

ench

-0.57

eno

-0.57

POSITIVE LOGITS

 Pulitzer

0.84

 sooner

0.84

Had

0.81

 someday

0.77

 Nobel

0.72

apo

0.71

 Wouldn

0.70

«

0.69

amation

0.68

MIT

0.65

Activations Density 0.689%

hypothetical situations or regrets in which actions could have been different

conditional statements and hypothetical scenarios involving outcomes

No Comments

No Known Activations