INDEX

Explanations

phrases indicating repetition or multiple occurrences of an action

oai_token-act-pair · gpt-3.5-turbo

occurrences of the word "times" 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

Reviewer

-0.93

tymology

-0.77

XT

-0.77

CHAT

-0.76

rt

-0.73

agra

-0.73

ourt

-0.73

aceous

-0.72

rats

-0.70

ogy

-0.69

POSITIVE LOGITS

 consecut

1.11

 throughout

0.76

cale

0.75

manship

0.74

fold

0.73

 apiece

0.71

soever

0.70

 concurrently

0.66

orial

0.65

pan

0.63

Activations Density 0.025%

phrases indicating repetition or multiple occurrences of an action

occurrences of the word "times" in various contexts

No Comments

No Known Activations