INDEX

Explanations

instances where a speculative or hypothetical outcome is being discussed

oai_token-act-pair · gpt-3.5-turbo

conditional phrasing and hypotheticals

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

wrote

-0.66

fighting

-0.63

action

-0.62

 Reporting

-0.59

Lind

-0.59

Is

-0.58

 Salman

-0.58

 touring

-0.57

Ren

-0.57

 Sale

-0.57

POSITIVE LOGITS

 suffice

1.08

ŃĶ

0.92

 overshadow

0.92

©¶æ

0.90

be

0.89

 imply

0.84

 conflic

0.83

 distort

0.83

 complicate

0.83

 alleviate

0.77

Activations Density 0.367%

instances where a speculative or hypothetical outcome is being discussed

conditional phrasing and hypotheticals

No Comments

No Known Activations