INDEX

Explanations

words related to action verbs and situational adjectives

oai_token-act-pair · gpt-3.5-turbo

phrases related to negative experiences or emotions

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

 inspecting

-0.67

 endeav

-0.63

 retaining

-0.61

 congratulated

-0.61

lication

-0.61

ledge

-0.61

 stances

-0.60

 Parables

-0.60

 positioning

-0.59

 browsing

-0.59

POSITIVE LOGITS

 fruition

0.78

 happening

0.78

 guiActiveUnfocused

0.75

 downhill

0.72

netflix

0.71

 undone

0.70

 ripple

0.70

ipples

0.69

 outweigh

0.66

 occurring

0.64

Activations Density 0.702%

words related to action verbs and situational adjectives

phrases related to negative experiences or emotions

No Comments

No Known Activations