INDEX

Explanations

the phrase "that" followed by a target or assertion

oai_token-act-pair · gpt-3.5-turbo

the word "that" in various contexts, indicating a focus on clauses or assertions within the text

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

onom

-0.63

stead

-0.62

Laughs

-0.61

atre

-0.61

oret

-0.60

onomic

-0.60

onsequ

-0.60

istant

-0.59

stal

-0.59

unction

-0.57

POSITIVE LOGITS

 accompanies

1.09

 justifies

0.89

 preceded

0.89

 arose

0.88

 existed

0.86

 they

0.86

 governs

0.82

 dominates

0.81

 exists

0.80

 prevailed

0.80

Activations Density 0.208%

the phrase "that" followed by a target or assertion

the word "that" in various contexts, indicating a focus on clauses or assertions within the text

No Comments

No Known Activations